Using the MATLAB artificial neural network toolbox to establish BP neural network model based on Bayesian regularization algorithm,a method of earthquake damage prediction for multistory masonry building based on the survey data of earthquake damage from earthquake events is proposed.The input earthquake damage factors include building story number,construction quality and house integrity.The output value is building＇s destructiveness under seismic action.The result shows that the method can predict earthquake damage of multistory masonry building with good achievement.